新疆北部地区积雪信息遥感反演研究

针对不同区域归一化雪被指数值(NDSI)上下浮动的问题,提出以对积雪响应敏感的NDSI和归一化植被指数(NDVI)为特征向量构建特征空间,建立基于MODIS数据的积雪丰度反演模型——NN模型。以高分辨率的Landsat TM数据对反演结果进行精度检验,NN模型反演积雪丰度值与获取自Landsat TM数据的NDSI值的相关系数R2为0.827。结果表明,通过建立NDSI—NDVI特征空间构建积雪丰度反演模型是可行的,且精度较高。因此,该方法对于了解干旱区积雪情况、评估其对新疆农业生产发展的影响具有一定现实意义。

Study on Remote Sensing Inversion of Snow Information in the Northern Region of Xinjiang

According to the normalized snow index(NDSI) fluctuate for different regions,the study put forward to construct feature space with the NDSI and vegetation index(NDVI) that sensitive in response to the snow,and establish the snow abundance inversion model(NN model) based on MODIS data.The precision inspection of inversion results conducted by the high-resolution Landsat TM data,the correlation coefficient R2 is 0.827 between NN model inversion abundance value with the NDSI calculated by Landsat TM data.The results showed that,through the establishment of NDSI-NDVI feature space to construct a snow abundance inversion model is feasible with high precision.Therefore,the method has a certain practical significance for understanding the arid area snow situation,assessing its impact on the development of agricultural production in Xinjiang.